I generally start my games with the visuals first, though I would not necessarily recommend this approach. It has the advantage of having nice looking assets at the start, but until you start developing, you are not going to necessarily know what you need or not. This can lead to wasted time on assets not used, assets needing reworking, etc.
Generally the reason I start on visuals first is time, especially in game jams or time-constrained projects, as often I can quickly hammer out something and "make it work" with the game mechanics. Generally this works out okay, as over time I have gotten a better idea of what I may need, but there are times where I have to... get creative. One reason I think this has worked for me is that I'm a bit of a jack-of-all trades, with a bent towards programming and 3D.
My advice to anyone looking to start a new project is this: Play to your strengths! Whatever you feel is your strength of what you like doing, try to make your project shine using that. Not only will it potentially help the game stand out, but you will likely enjoy the process more and that is important! Motivation is one of the biggest reasons I have seen that projects stall.
With that in mind, I would not recommend doing the inverse: leaving the parts you do not feel confident with to last. Leaving those parts to last will make it that much harder to finish the project, in my opinion, as they will continue to grow as other parts of the game progress. I feel its better to try and strike a balance between working on the parts you like and working on the parts you are not as confident on.
That's just my two cents though. :smile: